The role of digital tools in enhancing antiretroviral therapy adherence among adolescents and young adults with HIV: a systematic review of randomized controlled trials

Abstract

Background

Adolescents and Young Adults (AYA) face challenges in maintaining antiretroviral therapy (ART) adherence, necessitating the need for targeted and innovative interventions. Digital tools have shown potential in improving ART adherence following its integration into HIV care. However, there is paucity of comprehensive evidence on its usage and effectiveness in this population who exhibit higher levels of non-adherence to ART. Hence, we conducted a systematic review of randomized controlled trials (RCTs) to evaluate the effectiveness of digital tools in enhancing ART adherence among AYA.

Method

We searched multiple databases including Google Scholar, PubMed, Cochrane and Web of Science using appropriate keywords and Boolean operators. We screened articles for eligibility and ensured that only RCTs focusing on the role of digital tools in ART adherence among AYA with HIV were included.

Results

Of the 4,270 studies screened, 8 studies published between 2020 to 2024 met the eligibility criteria. The included studies involved study population between the ages of 10–24 years. The studies assessed the effectiveness of digital tools on ART adherence among AYA living with HIV with sample sizes ranging from 21 to 349 participants. The digital tools in the studies included SMS, gamified text messages, social media group, interactive voice response system, video conferencing, and a gamified mobile app. Only two studies showed statistically significant improvement in ART adherence. Adherence measures were found to be a key factor in influencing adherence outcomes as viral load-based measures were associated with the statistically significant studies.

Conclusion

Digital tools show potential in improving ART adherence among AYA living with HIV. Although only two studies reported statistically significant improvement in adherence, mHealth and SMS have shown promise as tools for improving ART adherence and achieving viral suppression. Future studies should employ viral load as a primary measure for the ART adherence as well as increase follow-up durations and sample size to ensure generalizability of results.

Introduction

The World Health Organization’s 2024 Epidemiological Fact Sheet estimates that approximately 39.9 million people were living with HIV globally by the end of 2023 [1] and out of this number, 3.4 million were youths aged 15–24 years. Adolescents and young adults (AYA) with HIV are a demographic that generally face challenges in maintaining adherence to antiretroviral therapy (ART) compared to other age groups [2]. Although there is a scarcity of age-categorized data on ART coverage, the available data indicates that there is a higher rate of discontinuity to HIV care and treatment adherence among AYA with HIV when compared to younger children or adults [3, 4]. This could be due to factors such as stigma and fear of disclosure to others, lack of social support and youth-specific services, and limited knowledge about HIV [5]. Studies conducted among adults, adolescents, and children, reveal that in other to obtain and maintain the full benefits of ART, an extremely high level of medication adherence is required [6, 7]. Consequently, AYA with HIV are generally more prone to risk of viral failure, secondary HIV transmission and drug resistance [8] compared to other age groups.

Digital tools have emerged as a promising solution to improving ART adherence and monitoring [9]. Digital tools are software, applications, or electronic devices that use digital technology to perform tasks, solve problems, or facilitate specific functions. In the context of ART adherence, digital tools refer to technology-based interventions designed to support individuals in consistently taking their anti-retroviral medications as prescribed by leveraging digital platforms and devices to address adherence barriers, provide reminders, and enhance patient engagement. They include mobile Health (mHealth), web-based interventions and smart devices [2]. The use of digital tools as an intervention to improve ART adherence have recorded great success in increasing retention among non-adherent patients and those that may be at risk of treatment interruptions [10]. A systematic review of interactive digital interventions reported that specific approches effectively enhances ART adherence and engagement in HIV care [2]. Similarly, mHealth interventions have shown postive outcomes in improving medication adherence and retention in care mong adolescents living with HIV, particularly in low-resource settings [11]. Additionally, mobile applications designed to promote engagement in HIV care and adherence among youth have demonstrated feasibility and acceptability in improving ART outcomes [12]. However, despite the recent increase and integration of digital tools into HIV care, their usage and effectiveness among AYA with HIV—who exhibit higher levels of non-adherence to ART—seems to be largely underreported. Hence, there is a need to synthesize current evidence to reflect the effectiveness of the use of digital tools in promoting adherence to ART among AYA with HIV. This systematic review aims to systematically determine the effect of digital tools in enhancing ART adherence among AYA with HIV from randomized controlled trials.

Methodology

Study protocol and registration

This systematic review was conducted following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [13]. The systematic review was registered on PROSPERO with registration number CRD42025636057.

The aim was to synthesize existing evidence on the role of digital tools in enhancing antiretroviral therapy (ART) adherence among adolescents and young adults living with HIV.

Search strategies

The literature search was conducted on 24th December 2024 in PubMed, Google Scholar (using Harzing’s Publish or Perish app), Cochrane and Web of Science using the following terms: (Digital tools) OR (Mobile applications)) OR (Telemedicine)) OR (eHealth)) OR (mHealth)) AND (Antiretroviral therapy adherence)) OR (ART adherence)) AND (Adolescents)) OR (Young Adults)) AND (HIV)) OR (People Living with HIV)) OR (PLHIV). These databases were selected as they index literature relevant to digital health interventions and HIV care and also based on public health research and RCTs. The reference list of systematic reviews and meta-analyses that were included in our initial search results were assessed for eligible articles, to capture additional peer-reviewed articles, which were included also. Institutional reports, conference proceedings, preprints, and non-peer reviewed articles were excluded to ensure that all included studies met rigorous peer-review standards. Identified articles were imported into EndNote to remove duplicates and then screened using Rayyan software.

Eligibility criteria

Studies were included if they met the following criteria:

• Participants involved only adolescents and young adults living with HIV ranging between the ages of 10–24 years as defined by the World Health Organization (WHO) [14]. While some studies defined AYA slightly differently (e.g., 13–24 years; 15–24 years), all included studies focused on individuals within the 10–24-year range (using WHO’s definition of AYA), ensuring relevance to our research question

• Evaluated digital tools (e.g., mobile health apps, SMS reminders) as an intervention.

• Included at least one quantitative outcome measure of ART adherence, such as self-reported adherence rates, viral suppression, or medication refill rates.

• Published in English in an available full text format between January 2014 and December 2024.

• Randomized controlled trials (RCTs) only.

• Had standard of care as a comparator

Exclusion criteria

Studies were excluded if they were not RCTs, or focused on other age groups, non-digital interventions, or did not report adherence-related outcomes.

Study selection

Two reviewers (MCA and COO) independently screened titles and abstracts for relevance, and conflicts were resolved by CEA. The process was repeated for full texts to identify potentially eligible studies by MCA and COO, and disagreements were resolved by CEA. All included RCTs were assessed for ethical compliance, and only studies that reported adherence to ethical research guidelines, such as obtaining informed consent and approval from relevant ethics committees, were included. As a result, ethical approval was not required for this systematic review. The studies that did not meet the eligibility criteria were excluded.

Data extraction and analysis

The following data were extracted into a Microsoft Excel spreadsheet by CEA: title, author, publication date, journal, study design, study population, study objectives, sample size and participants characteristics, digital tool intervention details, outcomes, key findings and conclusions.

Quality assessment

Quality assessment was performed using the Cochrane RoB 2 Excel sheet beta version 9. Two authors (MCA and COO) independently assessed each article for risk of bias and strength of evidence, resolving any disagreements through discussion and the intervention of CEA.

Figure 2 (shows the risk bias assessment for the included studies).

Fig. 2.

Quality assessment of the studies

Results

Initially, 5,429 studies were retrieved from databases and other methods (PubMed with 4200 articles, Google Scholar with 980, Scopus with 140, Web of Science with 97, and 12 articles from citation search). Title and abstract screened after duplicate articles (1,159 articles from all sources) were removed to see how they relate to the objective of the review. The screening was done based on the inclusion criteria as outlined in Sect. 2, following which 29 articles were retrieved for full-text screening. A total of 29 articles were given complete text examination and eligibility evaluation from which 8 articles were excluded based on ineligible population (demographic did not focus on adolescent and young adults living with HIV), 4 articles evaluated digital tools that were not designed to improve ART adherence, 6 articles were not RCTs, 2 articles did not include any digital tool intervention while 1 was a duplicate study. Finally, 8 articles were included in this systematic review.

Figure 1, a PRISMA flowchart provides a schematic diagram and summary of the selection process.

Fig. 1.

PRISMA 2020 flow diagram for new systematic reviews which included searches of databases and registers only

Characteristics of included studies

The included studies were published between 2020 to 2024; 2 articles in 2020 [15, 16], 3 in 2021 [8, 17, 18], 1 in 2022 [19], and 2 in 2024 [20, 21]. All included studies used randomized control trials to determine the effectiveness of the use of digital tools in promoting ART adherence among AYA living with HIV. Three studies were from Uganda [15, 17, 20] while 2 studies each were from Nigeria [8, 16] and the USA [18, 19], and 1 study was from Ghana [21]. Participants in all studies were adolescents and young adults ranging from 10–24 years as defined by the WHO [14]. The number of participants in each of the included studies ranged from 21 to 349, with 1 study focused on young men who have sex with men [18]. Interventions lasted between 22 to 94 weeks for all the studies. The study with the least sample size was a feasibility study and as such the study was preliminary research. Findings from small sample sizes may not be generalizable to the larger population, as they may not capture the diversity and heterogeneity of the broader group. Only 4 of the included studies had powered sample sizes to detect efficacy and the differences in efficacy [8, 16, 18, 20]. We included unpowered studies because they were pilot studies and can be helpful in the designing of fully-powered trials.

Study interventions and measures

Table 1 outlines the interventions (type of digital tools) and adherence measures utilized in the included studies.

Table 1.

Study and sample characteristics

Author & Year	Study Design	Study Objectives	Sample Size	Sample Characteristics
MacCarthy et al., 2020 [15]	Randomized Controlled Trial	To assess the acceptability, feasibility, and preliminary impact of SITA (SMS as an Incentive to Adhere) intervention to enhance antiretroviral therapy (ART) adherence among HIV-positive youth in Uganda	155	HIV-positive participants between the ages of 16–24 years, with at least 3 months of ART use. The participants were randomized into 3 groups: Standard of Care with no SMS intervention; Group T1 where participants received SMS messages only about their own adherence and Group T2 where participants received SMS messages about their own adherence as well as the adherence of their peers
Abiodun et al., 2021 [8]	Randomized Controlled Trial	To evaluate the feasibility, acceptability, and effectiveness of personalized and interactive SMS reminders in improving ART adherence among adolescents living with HIV in Southwest Nigeria	209	Participants within the ages of 15–24 years randomized into the either the intervention or standard of care
Dulli et al., 2020 [16]	Randomized Controlled Trial	To evaluate the effectiveness of a social media-based support group for ART adherence and retention in treatment	349	Participants aged between 15–24 years randomized to either the SMART group (Social Media to promote Adherence and Retention in Treatment) or Standard of Care
Twimukye et al., 2021 [17]	Randomized Controlled Trial	To evaluate the acceptability and feasibility of a mobile phone support tool, Call for Life Uganda (CFLU), in promoting adherence to ART among young adults living with HIV	21	Participants within the ages of 18–24 years randomized either to the intervention group or standard of care
Bwanika et al., 2024 [20]	Randomized Controlled Trial	To assess the effectiveness of the Call for Life Interactive Voice Response (CFL-IVR) mHealth tool in improving adherence to antiretroviral therapy (ART) and achieving viral suppression among young people living with HIV in Uganda	206	Participants between the ages of 15–24 years randomized to the intervention group or standard of care
Hightow-Weidman et al., 2021 [18]	Randomized Controlled Trial	To evaluate the effectiveness of a gamified mobile application, Epic Allies, in improving engagement in HIV care and ART adherence among young men who have sex with men	146	Males aged between 16–24 years randomized to either intervention or standard of care
Amico et al., 2022 [19]	Randomized Controlled Trial	To evaluate the efficacy of a 12-week triggered escalating real-time adherence (TERA) intervention with remote coaching and electronic dose monitoring (EDM) outreach for missed/delayed doses in improving adherence to antiretroviral therapy (ART) among youth living with HIV (YLWH) compared to the Standard of Care (SOC)	89	Participants between the ages of 13–24 years randomized into either TERA or SOC
Tarantino et al., 2024 [21]	Randomized Controlled Trial	To assess the feasibility, acceptability and preliminary efficacy of a Text-based Adherence Game (TAG) in improving ART adherence, viral suppression, and clinic attendance among young people living with HIV (YPWH) in Ghana	60	Participants between the ages of 18–24 where 30 were randomized into the TAG group while the other 30 were randomized into standard of care arm

SMS reminders

Three of the studies relied solely on SMS to deliver ART adherence support to their participants [8, 15, 17]. MacCarthy et al. incorporated two groups: one receiving individual SMS adherence feedback (T1) and the other receiving a combination of individual and peer adherence feedback (T2) [15]. Abiodun et al. and Twimukye et al. implemented a single treatment group that received interactive SMS reminders to support ART adherence and follow-up appointments [8, 17]. While adherence was assessed using electronic dose monitoring (EDM) and compared against a control group in MacCarthy et al.’s study [15], adherence was assessed through viral load, self-reports (Visual Analog Scale (VAS), AIDS Clinical Trial Group questionnaire), and pill counts in the latter studies [8, 17]. In Tarantino et al.’s study, the intervention arm received tailored, semi-automated, and gamified text messages throughout the intervention period [21], with adherence assessed through viral load and missed HIV clinic visits.

Social media

Dulli et al.’s study utilized a social media group – SMART Connections – to foster social support, ART adherence, and HIV-related knowledge [16]. Retention in HIV care was explored as an adherence outcome.

Mobile health app

In Bwanika et al.’s study, a mHealth tool integrated with an interactive voice response system, which offers daily pill reminders, clinic visit notifications, health tips, and a symptoms-reporting feature, was used as the intervention [20]. In this study, adherence was measured via viral suppression and retention in care. Epic Allies – a gamified mobile app intervention with features such as medication reminders, social support, and other self-management tools was used in Hightow-Weidman et al.’s study [18]. Adherence was equally measured using viral load.

Multiple digital tools (video conferencing, phone calls, text messages)

Amico et al. implemented a multilevel intervention that included coach-led video conferencing, phone calls, and text messages, and assessed adherence through viral load measurement [19].

ART adherence outcomes

Viral load suppression and retention in care

A summary of adherence outcomes is presented in Table 2. Only two studies reported statistically significant improvements in adherence associated with the intervention [8, 20]. In Abiodun et al.’s study, which was conducted in Nigeria, the mean difference in end-of-study (at 20 weeks) log10 viral load values between the control and intervention groups was 0.66 (95% CI: 0.26–1.06, p = 0.044) [8]. Additionally, the unadjusted odds ratio for achieving an undetectable viral load (≤ 20 copies/ml) was 1.356 (95% CI: 1.039–1.771, p = 0.001). However, the intervention did not influence self-reported measures of ART adherence including pill count [8]. Bwanika et al.'s study revealed that viral suppression at 12 months was 73.6% (67/91) in the intervention arm, compared to 51.9% (40/77) in the standard of care arm (p < 0.01) [20]. Additionally, retention in care at 12 months was 88.4% (91/103) in the intervention arm versus 74.7% (77/103) in the standard of care arm (p = 0.01) [20].

Table 2.

Intervention details, outcomes and findings

Author & Year	INTERVENTION				ART Adherence Outcome (Primary or Secondary)	Other Outcomes	Findings	Statistical results	Conclusion
	Means	Frequency	Mode of Delivery (Synchronous Or Asynchronous)	Length
MacCarthy et al., 2020 [15]	SITA (SMS as an Incentive to Adhere) was used which provided feedback on adherence. The feedback was either about personal adherence or involving both personal adherence and the average adherence of peers	Weekly	Asynchronous	36	Primary	Acceptability and feasibility of the SMS-based intervention, including participant engagement and retention	Providing self-feedback on adherence (T1) did not significantly improve adherence compared to standard care. However, integrating peer-comparison (T2) showed a higher and a numerically steady increase over time	T1 vs. Control: −3.8 percentage points (pp) (95% CI: −9.9, 2.3) T2 vs. Control: + 2.4 pp (95% CI: −3.0, 7.9) For T1: Negative treatment effects observed in 3 out of 4 weeks intervals For T2: Initial 9 weeks: + 3 pp increase in adherence Final 9 weeks: + 9 pp increase in adherence	The findings demonstrate that peer comparison (T2) is a more effective approach for enhancing ART adherence than self-feedback (T1) or standard care as it leverages social motivation to achieve improvements in adherence
Abiodun et al., 2021 [8]	Personalized and interactive SMS reminders sent to adolescents living with HIV	Daily SMS reminders and biweekly SMS notifications for clinic appointments	Asynchronous	52	Primary	Feasibility and acceptability of the SMS delivery	No significant difference was found between the intervention and control groups in self-reported adherence	1. VAS (continuous): Risk difference = −2.050 [−5.520—1.420]; p-value = 0.246 2. VAS (adherence ≥ 95%): Risk ratio = 1.201 [0.912—1.581]; p-value = 0.189 3. Viral load (≤ 20 copies/ml): Risk ratio = 1.356 [1.039—1.771]; p-value = 0.022 3b. Log Viral Load: Risk difference = 0.660 [0.26—1.06]; p-value = 0.001 4. Pill count: Risk difference = −0.034 [−0.104—0.037]; p-value = 0.347 −0.034 [−0.104—0.037] 5. AIDS Clinical Trials Group questionnaire scores: Risk difference = −0.025 [−0.068—0.019]; p-value = 0.261	SMS reminders may enhance viral suppression and therefore improve outcomes in adolescent HIV The impact of tailored SMS reminder on self-reported adherence measures was inconclusive, highlighting the need for better adherence assessment methods
Dulli et al., 2020 [16]	The Social Media-based intervention (SMART Connections) used secret Facebook groups to deliver content related to HIV care, adherence, and social support. This was facilitated by trained peers also living with HIV	Nearly daily activities within the Facebook group over a 22-week period	Primarily asynchronous	22	Secondary	Retention in HIV care, HIV-related knowledge, and social support	Retention in care did not differ significantly between intervention and control groups. The likelihood of staying in care without a gap exceeding 28 days was slightly higher in the intervention group compared to the control group at most time points, except at 120 days. However, the 95% confidence intervals for both groups overlapped at all time points Also, no significant differences were observed in ART adherence. However, HIV-related knowledge significantly improved in the intervention group compared to the control group	Probability of being retained in care (Intervention [95% CI] vs. Control [95% CI]): 0 day: (0.99 [0.96 – 1.00] vs. 0.98 [0.94 – 0.99]) 60 days: (0.81 [0.73 – 0.86] vs. 0.73 [ 0.65 – 0.79]) 90 days: (0.73 [0.65 – 0.79] vs. 0.70 [0.62 – 0.77]) 180 days: 0.52 [0.43 – 0.60] vs. 0.53 [0.45 – 0.61]) 270 days: (0.50 [0.41 – 0.58] vs. 0.45 [0.36 – 0.54]) Adherence: (χ2 = 0.32; p-value = 0.57)	The intervention improved HIV knowledge and had high acceptability but did not significantly impact retention or adherence
Twimukye et al., 2021 [17]	The Call for Life Uganda (CFLU) system delivered interactive voice response (IVR) or SMS-based daily pill reminders, weekly health tips, and clinic appointment reminders. It also allowed participants to report symptoms	Daily reminders for pills, weekly health tips, and appointment notifications as scheduled	Both synchronous and asynchronous	52	Primary	Management of stigma, psychosocial support	Participants reported improved medication adherence and reduced forgetfulness due to pill reminders Additionally, the tool provided health tips to address HIV-related stigma and supported privacy through PIN-protected interactions	N/A	The CFLU system was acceptable and feasible for improving ART adherence in resource-limited settings, despite technical challenges
Bwanika et al., 2024 [20]	The Call for Life Interactive Voice Response (CFL-IVR) tool provided tailored daily pill reminders, clinic visit notifications, and health tips, with an option for symptom reporting	Daily pill reminders, weekly health messages, and clinic appointment reminders every six months	Both synchronous and asynchronous	94	Secondary (Supports the primary outcomes of viral suppression 12 and 6 months and retention in care)	Retention in care at 6 and 12 months, viral suppression at 6 months	At 12 months, viral suppression was significantly higher in the intervention group compared to the control group There was no statistically significant difference between the study arms in secondary outcomes, including viral suppression at month 6 Retention in care showed a statistically significant difference between groups, with the intervention group having an 8.7% higher retention rate at 6 months (95% CI: 0.2% to 17.3%) and a 13.5% higher retention rate at 12 months (95% CI: 3.1% to 24.0%) Participants in the control group reported higher rates of missed pills due to forgetfulness or being away from home compared to the intervention group	Viral suppression at 12 months; ≥ 1000 copies/ml: Diff. in percentage points = −0.20 (−0.35 – 0.06); p-value = < 0.01 Viral suppression at 6 months; ≤ 1000 copies/ml: Diff. in percentage points = −0.09 (−0.23—0.06); p-value = 0.23 Retention at 12 months: Diff in pp = 13.5% (3.1% to 24.0%); p-value < 0.01 Retention at 6 months: Diff in pp = 8.7% (0.2% to 17.3%); p-value = 0.05	The CFL-IVR tool was effective in improving both ART adherence and retention in care among young people living with HIV
Hightow-Weidman et al., 2021 [18]	The Epic Allies mobile application, which incorporated features such as medication reminders, gamified rewards for adherence, Self-monitoring tools to track progress and Tailored feedback based on user behaviour, all designed to improve adherence to ART	Participants were instructed to use the app daily over 26 weeks	Asynchronous	39	Secondary	Viral load suppression, Engagement in care, ART uptake and durability of viral suppression after the intervention	Both intervention and control groups showed improved viral suppression at 26 weeks. Also, engagement rates in care and ART adherence improved but did not significantly differ between the two groups	1. Viral suppression a. At 13 weeks: Adjusted Risk Ratio, ARR = (1.17 [0.87–1.57]; p value (unadjusted) = 0.310) b. At 26 weeks: ARR = (0.93 [0.73–1.18]; p-value (unadjusted) = 0.52) c. At 39 weeks: ARR = (1.02 [0.78–1.35]; p-value (unadjusted) = 0.869) 2. ART uptake a. At 13 weeks: ARR = (0.73 [0.57—0.93]; p-value (unadjusted) = 0.011) b. At 26 weeks: ARR = (0.90 [0.78—1.03]; p-value (unadjusted) = 0.115) c. At 39 weeks: ARR =  (0.73 [0.51—1.04]; p-value (unadjusted) = 0.082) 3. Adherence (≥ 90%) a. At 0-week ARR =  (0.75 [0.49- 1.14]; p-value (unadjusted) = 0.183) b. At 13-week ARR =  (0.88 [0.66—1.16]; p-value (unadjusted) = 0.353) c. At 26-week ARR =  (1.54 [1.03- 2.29]; p-value (unadjusted) = 0.035) d. At 39-week ARR =  (1.03 [0.72—1.48]; p-value (unadjusted) = 0.877)	The app showed potential for enhancing ART adherence and viral suppression particularly among regular users. However, challenges such as decreased app usage overtime and small sample size restrict generalizability. Further studies should ensure optimized app engagement
Amico et al., 2022 [19]	The intervention was the Triggered Escalating Real-time Adherence (TERA) system, which combined electronic dose monitoring (EDM) pill bottles and remote coaching. The system provided real-time feedback on adherence and allowed outreach by coaches when doses were missed 41 (21, 59) 72 (47, 89)	Daily pill monitoring with coaching sessions as needed during the 12-week period	Both synchronous and asynchronous	48	Secondary	Viral Suppression (HIV-1 RNA < 200 copies/mL at 12 weeks)	Median percent days with electronic dose monitoring opening was higher in the intervention group (72%) compared to standard care (41%) However, there was no significant difference in viral suppression between intervention and control groups at 12 weeks (TERA: 35%; SOC: 36%)	1. Viral suppression (HIV-1 RNA < 200 copies/mL at 12 ± 2 weeks): Diff in pp (complete case) = (− 0.3 [− 24.6 – 23.1]; p > 0.99 2a. Adherence (% doses taken by 12 weeks): 0–12 weeks SOC vs. TERA (Median [Q1, Q3]) = 41 (21, 59) vs. 72 (47, 89); p-value < 0.001 2b. GAP incidence rate (number of 7 day gaps (defined as 7 consecutive days without opening the EDM) over 12 weeks: TERA to SOC IR ratio = (0.40 [0.30—0.53], p-value < 0.001	The TERA intervention improved adherence but did not significantly impact viral suppression
Tarantino et al., 2024 [21]	TAG (Text-Based Adherence Game) was delivered through SMS and involved detective-themed gamification storyline, personalized messages on adherence, a point leaderboard system and opportunities to request social support from clinic staff or peer educators	Daily messages in month 1; every other day during months 2 and 3; every three days during months 4 and 5; weekly in month 6	Asynchronous	52	Primary	Viral load and missed HIV clinic visits	Participants in the TAG arm showed significant improvement in ART adherence from baseline to 12-month follow-up, while SOC participants experienced a decline	1. ART adherence for 12-month follow-up (F, p-value): a. 3-item scale (past 30 days) = (0.16, 0.85) b. Visual Analog Scale (past 28 days) = (3.65, 0.03) c. Log viral load = (0.01, 0.99) d. Missed HIV clinic visits (≥ 1) = (0.37, 0.69)	TAG intervention showed a promise in improving ART adherence

No statistically significant adherence-related findings were reported in the other studies [16–20, 22]. In MacCarthy et al.’s study, the treatment 2 (T2) group, which received both individual and peer adherence feedback, showed a slight improvement in adherence, though this result was not statistically significant. After controlling for baseline adherence, the T2 group had 2.4% higher adherence compared to the control group (95% CI: −3.0 to 7.9), while the T1 group had 3.8% lower adherence than the control group (95% CI: −9.9 to 2.3) [15]. The authors did acknowledge that the pilot study was not designed to detect statistical significance. Dulli et al. also reported no significant difference in self-reported adherence between the intervention and control groups (p = 0.57) [16]. However, HIV-related knowledge was significantly higher in the intervention group than in the control group at endline (p = 0.003) [16]. In Twimukye et al.’s study, AYA living with HIV reported improvement in medication adherence. Qualitative analysis of participant interviews confirmed that the CFLU tool improved adherence to routine clinical appointments, thereby supporting ART adherence. Hightow-Weidman et al. reported a notable increase in the proportion of participants achieving viral suppression in both the intervention group (61.0%) and the control group (54.5%) at 13 weeks; however, the difference was not statistically significant (Adjusted Risk Ratio, ARR = 1.17; 95% CI: 0.87–1.57; p-value = 0.310) [18]. Moreover, at 26 weeks, viral suppression in the intervention arm remained relatively stable at 62.9%, while it increased to 73.5% in the control arm (ARR = 0.93; 95% CI 0.73–1.18) [18]. Also, viral suppression stabilized at approximately 64% in both groups, at 39 weeks (ARR = 1.02, 95% CI 0.78–1.35) [18].

Amico et al. reported that virologic success (defined as HIV-1 RNA < 200 copies/ml measured between weeks 10–14) was achieved in 16 of 45 participants (36%; 95% CI: 22%, 51%) in the SOC arm and in 15 of 43 participants (35%; 95% CI: 21%, 51%) in the TERA arm, however this was not statistically significant (p > 0.99) [19].

Finally, Tarantino et al. found no significant differences in viral suppression rates between the groups at 6 months, with 45% in the TAG group and 48% in the SOC group (p = 0.98) [21]. Also, the intervention had no significant influence on missed HIV clinic visits (p = 0.70) [21].

Quality assessment

Quality assessment was performed using the Cochrane Risk of Bias 2 Excel sheet beta version 9. From the 8 studies assessed for bias, a low risk of bias was found in the studies [8, 15, 18–21], as the processes involved were well stated. Some concerns, however, were found in [15, 16, 19]. Assessment from Twimukye et al. [17] showed a high risk of bias due to the randomization process and a greater portion of the participants being in the intervention arm. Moreover, deviation from the intended intervention also negatively impacted the quality of the study. Quality assessments for studies [16–20] were quite concerning in different domains such as randomization process, intended intervention, outcome measurement, outcome data, and selection of the reported result.

Figure 2 shows the risk bias assessment for the included studies.

Discussion

This systematic review explored the roles of digital tools in improving ART adherence among AYA living with HIV. A review of RCTs was done because RCT has demonstrated the highest reliability in assessing the effectiveness of interventions and has been regarded as the gold standard for evaluating intervention outcomes over the past ten years [22]. Its suitability in providing answers to clinical questions regarding the evaluation of diverse treatments solely lies in its propensity to minimize bias [23]. In addition, to assess the effectiveness of an intervention, it may be necessary to examine it for a significant period to determine its meaningful impact. Half of the included studies in this review were observed for a minimum of one year [8, 17, 20, 21], the remaining studies were mostly 11 months and below [15, 16, 18, 19]. Regardless of the intervention length, long-term authenticity is not guaranteed in studies spanning more than 12 months as indicated in one study [18] where a relatively stable intervention effect at 26 weeks was observed. This highlights that the short-term effects of interventions may not be applicable in the long run, as their benefits may not be sustained. Furthermore, since ART requires lifelong adherence to achieve optimal viral suppression, interventions with limited effectiveness may have a minimal impact on treatment success. Therefore, to confirm the effectiveness of an intervention, future studies may need to assess its impact over a longer period to ensure lasting results [24].

Overall, two studies demonstrated statistically significant improvements in ART adherence attributable to the intervention, whereas six studies did not. This aligns with the findings of a systematic review by Griffee et al. (2022) where the effectiveness of digital interventions in improving ART adherence among youth living with HIV in sub-saharan Africa was accessed [2]. There, only two of the included studies found significant intervention-related improvement in viral suppression. The study found that SMS-based interventions led to modest but significant improvements in ART adherence, consistent with the findings of Abiodun et al. in this review [8]. Similarly, the effectiveness of technology-based interventions in improving adherence has been demonstrated in other health conditions, such as tuberculosis, where a meta-analysis found that digital tools significantly enhanced treatment adherence, completion, and success [25]. These findings highlight the growing role of digital interventions in supporting long-term treatment outcomes. Notably, all the included studies were published within the last 4 years, reiterating the increasing interest in digital tools for ART adherence among AYA living with HIV. This further underlines the need for further investigation in this area.

Notwithstanding the limited evidence available to support the significant role of these digital tools, some notable observations merit further consideration [15, 16]. Viral load-based adherence measures are considered the "gold standard" for evaluating HIV treatment [26], thereby supporting the findings of the two studies that showed statistically significant adherence related to viral load [8, 20]. Other studies employed electronic or self-reported measures [15, 16], which some sources have noted offer less precise estimates of adherence. Self-reported adherence, though prone to overestimation, remains a widely accepted method due to its simplicity, and good specificity [27, 28]. A review by Simoni et al. further highlights the reliability of self-reported adherence while acknowledging its limitations [29]. Although the review reports that self-reported adherence correlates significantly with indirect adherence measures, viral load, and CD4 count, it also points out inconsistencies due to variations in recall periods and response formats. One drawback of self-reported adherence to ART is the recall bias and the potential for patients to overestimate or underestimate their adherence behaviour as they may forget missed doses or misunderstand the meaning of adherence. Also, social desirability bias is another shortcoming of self-reported adherence measures. Despite these drawbacks, the self-questionnaire is a widely used method for assessing adherence in HIV patients. It is simple, adaptable, quick, and cost-effective, despite the limitations of overestimation and recall biases associated with it. The design and complexity of the questionnaire can greatly impact the study outcomes [24]. Additionally, self-reported measures may not be suitable for individuals with cognitive impairments. This emphasizes the need for standardization in adherence assessment to enhance accuracy.

Factors such as small sample sizes, insufficient high-network coverage, and loss to follow-up may be key reasons why studies using viral-load adherence measures did not show statistically significant improvements in ART adherence [18, 19, 21]. Specifically, Dulli et al. highlighted the absence of a viral load measure as a limitation in their study designs [16]. Even though a significant association was found between viral suppression and adherence per EDM in Amico et al.’s study [19], Abiodun et al. attribute the absence of a correlation between viral load and pill count, visual analog scale (VAS), or AIDS Clinical Trials Group (ACTG) questionnaire results to the limitations of these three measures, including the common accumulation of leftover pills and the impact of forgetfulness and social desirability on self-reported data [8]. Future research should take into account viral load as a key measure to more precisely assess the impact of digital interventions on ART adherence.

The findings from three studies suggest that text message reminders may help enhance adherence to ART [8, 15, 17] although only one study found statistically significant improvement in ART adherence [8]. Mobile phone interventions are cost-effective and convenient for sending SMS to a large number of potential users across a broad region [30]. This approach could play a crucial role in advancing the UNAIDS 95–95-95 strategy and providing comprehensive ART adherence support. Using text message interventions to enhance ART adherence could be vital in achieving the second and third 95 s of the UNAIDS 95–95-95 initiative aimed at ending the AIDS epidemic. Also, with the majority of the included studies set in resource-limited settings like sub-Saharan Africa, SMS-based adherence interventions tend to be more effective in these areas where basic mobile phones are more common than smartphones, and where text messaging is more widespread than social media or internet usage [31].

Contrasting results seem to present over the effectiveness of weekly SMS reminders over daily SMS reminders. In this systematic review, the significant improvement in ART adherence as seen in Abiodun et al.’s study was attributed to daily SMS reminders rather than weekly reminders [8]. A previous review highlighted that weekly SMS messages enhanced adherence, while daily SMS messages did not show the same effect [28]. This finding aligns with a recent systematic review and meta-analysis by Shah et al., which reported that while SMS-based interventions showed some benefit, their impact was highly variable depending on message frequency, personalization, and user engagement [32]. Interestingly, this meta-analysis also found that weekly SMS reminders were more effective than daily reminders, whereas previous reviews have reported conflicting results on the frequency of SMS interventions [28]. Although participants may become frustrated with receiving frequent SMS reminders, it is well-established that repetitive actions can eventually form habits.

Furthermore, the 2-way text messaging feature, where participants were asked to respond to SMS messages, is a promising aspect of the SMS-based intervention that plays a key role in enhancing ART adherence. This was particularly noteworthy in the review, as Abiodun et al.'s study incorporated this feature into their intervention design, further supporting the inclusion of such elements in the development of SMS-based interventions [33, 34].

Another notable result in this review showed that mHealth tools can be effective in significantly improving ART adherence among AYA living with HIV [20]. The IVRS incorporated into the mHealth tool used in Bwanika et al.'s study demonstrated significant viral suppression, emphasizing the beneficial impact of mHealth tools on key outcomes such as ART adherence and retention [35]. The confidentiality and privacy provided by the IVRS feature, protected by a 4-PIN code, make this intervention particularly attractive to AYA living with HIV, ensuring strong ART adherence and retention in HIV care.

Despite the potential demonstrated by digital tools in improving adherence to ART, some challenges such as unreliable internet connectivity, limited access to digital tools, poor digital literacy, and cost present as barriers limiting the reach and the use of these interventions especially in low-resource settings [36]. Future studies should evaluate strategies to address these barriers such as developing low-cost digital health programs, and integrating digital literacy in HIV care.

Implications for practice and policy

This study brings to light the potential of digital interventions in improving ART adherence among adolescents and young adults (AYA) living with HIV. Hence, healthcare providers should consider integrating digital tools such as SMS reminders, interactive voice response systems (IVRS), and mobile health (mHealth) applications into routine HIV care, particularly in resource-limited settings where mobile phones are widely used [30]. Additionally, since it has been shown that two-way SMS messaging is linked to improved adherence, interventions should focus on prioritizing interactive features to promote engagement rather than relying solely on one-way reminders [8]. It is also important that healthcare professionals consider viral load monitoring as the key adherence measure, as self-reported adherence—though widely used—may lead to overestimation and recall bias [27, 28]. Adherence measurement should also be strengthened by objective viral load tracking as this will provide a more accurate assessment of intervention impact [8, 20].

In order to fully maximize the benefits of digital health interventions, national HIV treatment programs should integrate these interventions into policies and ensure that there are guidelines that will address ethical concerns such as data security, and patient confidentiality. Policymakers must also address barriers to accessibility by expanding mobile and internet coverage in unreached areas, as limited connectivity remains a challenge to the widespread adoption of mHealth solutions [31]. Funding agencies should also strive to give precedence to long-term studies to ensure intervention sustainability, as short-term effectiveness may not translate into lasting adherence improvements. The predominance of studies from sub-Saharan Africa and the United States reiterates the need for research in other underrepresented regions such as Asia and Latin America in order to improve the global applicability of these digital interventions. Ultimately, incorporating digital tools into practice and policy will enable stakeholders to enhance adherence support, improve treatment outcomes, and contribute to achieving the UNAIDS 95–95-95 goals for HIV epidemic control.

Limitations

With only eight studies included, this review restricts our ability to draw definitive conclusions about the effectiveness of digital interventions in enhancing ART adherence among AYA living with HIV. This limitation makes it challenging to generalize our findings. However, a good number of digital tools were featured in the review, thus presenting a solid validation of the conclusion. Another limitation of this review is the geographical concentration of the included studies, which may result in selection bias with most conducted in Sub-Saharan Africa (Uganda, Ghana, Nigeria) and the United States. While this gives an idea of the digital interventions employed in high-burden and resource-limited settings, the lack of data from other regions — such as Asia and Latin America — raises concerns about the global applicability of these findings. Moreso as digital health interventions may be influenced by some factors including sociocultural norms, health systems structures, and technological accessibility, all of which would vary significantly in different regions. Other potential biases include recall and social desirability biases; since most of the adherence measures were based on self-reported outcomes, errors are bound to occur.

Only one included study sampled a defined key population, MSM, thereby limiting the generalization of our findings to the broader population of AYA living with HIV. Some of the included studies had small sample sizes, which may have compromised the validity of the findings and reduced both generalizability and statistical power. A sample size smaller than ideal can lead to inaccurate results and conclusions about the impact of SMS reminders. Nevertheless, the straightforward SMS reminder played a crucial role in maintaining optimal ART adherence. Additionally, this study included only RCTs. While RCTs are considered the gold standard in providing information about the effectiveness of interventions, they may not capture real-world implementation challenges or patient preferences like observational or qualitative studies. Future studies should consider using broader inclusion criteria to capture findings from non-RCTs and gray literature enhancing a more comprehensive understanding of the role of digital tools in ART adherence among AYA living with HIV. Finally, differences in adherence measures across studies may introduce measurement bias, affecting the comparison and aggregation of results.

Conclusion

Digital tools show potential in improving ART adherence among AYA living with HIV. Although only two studies reported statistically significant improvement in adherence, this underlines the promise of mHealth and SMS as tools for improving ART adherence and achieving viral suppression. SMS offers a cost-effective approach and ease of use especially for low-resource settings where internet connections may be weak. Factors such as frequency of reminders and the inclusion of interactive features play vital roles in the effectiveness of these interventions. The use of different adherence measures may have led to the non-significant findings of majority of the studies. Future studies should employ viral load as a key measure for a more precise assessment on the impact on ART adherence. Also, larger sample size and longer follow up durations would ensure the accuracy and generalizability of results.

Abbreviations

AYA

Adolescents and Young Adults

HIV

Human Immunodeficiency Virus

ART

Antiretroviral Therapy

UNAIDS

Joint United Nations Programme on HIV/AIDS

mHEALTH

Mobile Health

Authors' contributions

CEA conceptualized the study; All authors were involved in the literature review; CEA & IJO extracted the data from the reviewed studies; CEA prepared Fig. 1; OA prepared Fig. 2. CEA prepared Tables 1 & 2. All authors (CEA, MCA, IJO, UEC, COO, OA, VOA, CSA, ISB, KEN, AI) wrote the final and first drafts. All authors (CEA, MCA, IJO, UEC, COO, OA, VOA, CSA, ISB, KEN, AI) read and approved the final manuscript.

Funding

No funding was received for this study.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.